To Treat or Not to Treat

Brief, asymptomatic atrial fibrillation seven years later

For many doctors in academic medicine, publishing in peer reviewed journal is an important part of their careers. Fortunately, over the last 25 years, promotions criteria have matured so it is not quite publish or perish anymore. People can rise through the academic ranks as gifted clinicians, creative and effective educators, and talented administrators.[i]

Though I’ve never been a traditional researcher, I’ve added quite a bit to “medical literature” over the years. Being perfectly honest, not all the articles have aged well. I am actually embarrassed by a few of them and many are forgettable. There are some, however, that I remain proud of. One of these was an article that Vinay Prasad and I wrote back in 2016: Wearables, Smartphones and Novel Anticoagulants: We Will Treat More Atrial Fibrillation, but Will Patients Be Better Off?

The idea behind the article was that we were heading into an era when atrial fibrillation was going to be easier and easier to detect in people who had no reason to be screened. A healthy young person might buy an apple watch so she could listen to podcasts on her runs and then, weeks later, learn that she had brief, asymptomatic episodes of atrial fibrillation. At the same time, anticoagulation was getting safer and easier. Our concern expressed in the article was that atrial fibrillation screening might become the norm, by default, and although screening might be beneficial, there was a reasonable likelihood that its harms would outweigh its benefits.

Last week, the ARTESIA study was published in the NEJM. This study made me happy for two reasons. First, it renewed my faith that occasionally well-done studies address interesting and important questions. Second, the results supported our suspicions from seven years ago that screening might do more harm than good.

Before I get to my analysis, I admit that I am honing in on Dr. Mandrola’s turf here writing about an atrial fibrillation trial. He actually discussed this trial a couple of weeks ago on Medscape.

ARTESIA did not enroll my Apple Watch wearer from above -- I’ll get to that in a bit. The design of ARTESIA was to take patients who were already having their hearts monitored (they had pacemakers or ICDs) and were found to have subclinical atrial fibrillation lasting six minutes to 24 hours. Patients were then randomly assigned to apixaban or 81 mg of ASA. The primary outcome was stroke or systemic embolism and the primary safety outcome was major bleeding.

4012 patients were included in the analysis. The mean age was about 77 years old and the patients were at high risk of stroke from atrial fibrillation with a mean CHA₂DS₂-VASc score of nearly 4. (Generally scores of 2 in men or 3 in women indicate anticoagulation) After a mean follow-up of 3.5 years, anticoagulation reduced the rate of stroke or systemic embolism by 1.7% (HR: 0.63; 0.45 to 0.88; P=0.007). On the other hand, major bleeding occurred in 1.3% more patients in the anticoagulation group (HR: 1.80; 95% CI, 1.26 to 2.57; P=0.001). This difference was driven by GI bleeding. There were no differences between the two groups in terms of fatal bleeding or symptomatic intracranial hemorrhage.

So, what does this all mean. From the perspective of the study population, the results are a wash. A patient who already has a heart monitor in place and is found to have brief, asymptomatic atrial fibrillation might choose to go on anticoagulation or aspirin.[ii]

The study is even less telling for our healthy young woman with her new apple watch. Her risk of stroke would be much lower, as would her risk of complications of anticoagulation. One could argue that finding atrial fibrillation would benefit her by bringing her to medical attention and being evaluated for hyperthyroidism. Of course, being “brought to medical attention” is not always a good thing.

I’ll end this piece in the same way Vinay and I finished our article in 2016. “The stars have aligned to search for and treat more atrial fibrillation; however, it remains uncertain whether or not our patients will be better off.”

---

[i] I’m still hoping for a time that publishing thoughtful articles on Substack counts for something.

[ii] About 60% of these patients were on an antiplatelet drug at the start of the trial so the decision would most commonly be to remain on aspirin or switch to a XA inhibitor.